Investigation of multi level linear, general linear and nonlinear growth models and application in livestock field

In this study, it is aimed to evaluate growth curves which have an important place in the field of animal husbandry by using general linear regression, nonlinear regression and two level linear regression models and to find the best explaining model of growth. For this purpose, the individual body weight values of 74 male broiler chickens were taken from hatching up to sixth week. Data were analyzed with models that respectively, Bertalanffy, Gompertz, Brody, Logistic, linear, cubic, unconditional linear growth model with random slope (Model I), the unconditional linear growth model with random slope and random intercept (Model II), and the unconditional quadratic growth model with random slope and random intercept (Model III). The coefficient of determination, mean square error, Log likelihood (II), Akaiki Information Criterion (AIC) Corrected AIC (AICC) and Bayesian Information Criterion (BIC) were used to find the model that best describes growth. At the end of the study, Bertalanffy, Gompertz, Brody and Logistic models could only be compared with themselves because they used average body weight values. In nonlinear growth models, Gompertz was the best model for explaining growth for study data. Linear regression models are not preferred models because they use average live weight values. Of the two-level linear regression models, Model III, which has the smallest -2II, AIC, AICC and BIC values, was the best explaining growth.